def __init__(self, user):
self.user = user
self.resource_manager_api = None
self.mapreduce_api = None
self.history_server_api = None
self.spark_history_server_api = None
if list(YARN_CLUSTERS.keys()):
self.resource_manager_api = resource_manager_api.get_resource_manager(user.username)
self.mapreduce_api = mapreduce_api.get_mapreduce_api(user.username)
self.history_server_api = history_server_api.get_history_server_api(user.username)
self.spark_history_server_api = spark_history_server_api.get_history_server_api() # Spark HS does not support setuser
def test_update_properties(self):
finish = []
finish.append(MR_CLUSTERS.set_for_testing({'default': {}}))
finish.append(MR_CLUSTERS['default'].SUBMIT_TO.set_for_testing(True))
finish.append(YARN_CLUSTERS.set_for_testing({'default': {}}))
finish.append(YARN_CLUSTERS['default'].SUBMIT_TO.set_for_testing(True))
try:
properties = {
'user.name': 'hue',
'test.1': 'http://localhost/test?test1=test&test2=test',
'nameNode': 'hdfs://curacao:8020',
'jobTracker': 'jtaddress',
'security_enabled': False
}
final_properties = properties.copy()
submission = Submission(None, properties=properties, oozie_id='test', fs=MockFs())
assert_equal(properties, submission.properties)
submission._update_properties('jtaddress', 'deployment-directory')
assert_equal(final_properties, submission.properties)
cluster.clear_caches()
fs = cluster.get_hdfs()
jt = cluster.get_next_ha_mrcluster()[1]
final_properties = properties.copy()
final_properties.update({
'jobTracker': 'jtaddress',
'nameNode': fs.fs_defaultfs
})
submission = Submission(None, properties=properties, oozie_id='test', fs=fs, jt=jt)
assert_equal(properties, submission.properties)
submission._update_properties('jtaddress', 'deployment-directory')
assert_equal(final_properties, submission.properties)
finish.append(HDFS_CLUSTERS['default'].LOGICAL_NAME.set_for_testing('namenode'))
finish.append(MR_CLUSTERS['default'].LOGICAL_NAME.set_for_testing('jobtracker'))
cluster.clear_caches()
fs = cluster.get_hdfs()
jt = cluster.get_next_ha_mrcluster()[1]
final_properties = properties.copy()
final_properties.update({
'jobTracker': 'jobtracker',
'nameNode': 'namenode'
})
submission = Submission(None, properties=properties, oozie_id='test', fs=fs, jt=jt)
assert_equal(properties, submission.properties)
submission._update_properties('jtaddress', 'deployment-directory')
assert_equal(final_properties, submission.properties)
finally:
cluster.clear_caches()
for reset in finish:
reset()
def test_update_properties(self):
finish = []
finish.append(MR_CLUSTERS.set_for_testing({'default': {}}))
finish.append(MR_CLUSTERS['default'].SUBMIT_TO.set_for_testing(True))
finish.append(YARN_CLUSTERS.set_for_testing({'default': {}}))
finish.append(YARN_CLUSTERS['default'].SUBMIT_TO.set_for_testing(True))
try:
properties = {
'user.name': 'hue',
'test.1': 'http://localhost/test?test1=test&test2=test',
'nameNode': 'hdfs://curacao:8020',
'jobTracker': 'jtaddress',
'security_enabled': False
}
final_properties = properties.copy()
submission = Submission(None, properties=properties, oozie_id='test', fs=MockFs())
assert_equal(properties, submission.properties)
submission._update_properties('jtaddress', 'deployment-directory')
assert_equal(final_properties, submission.properties)
cluster.clear_caches()
fs = cluster.get_hdfs()
jt = cluster.get_next_ha_mrcluster()[1]
final_properties = properties.copy()
final_properties.update({
'jobTracker': 'jtaddress',
'nameNode': fs.fs_defaultfs
})
submission = Submission(None, properties=properties, oozie_id='test', fs=fs, jt=jt)
assert_equal(properties, submission.properties)
submission._update_properties('jtaddress', 'deployment-directory')
assert_equal(final_properties, submission.properties)
finish.append(HDFS_CLUSTERS['default'].LOGICAL_NAME.set_for_testing('namenode'))
finish.append(MR_CLUSTERS['default'].LOGICAL_NAME.set_for_testing('jobtracker'))
cluster.clear_caches()
fs = cluster.get_hdfs()
jt = cluster.get_next_ha_mrcluster()[1]
final_properties = properties.copy()
final_properties.update({
'jobTracker': 'jobtracker',
'nameNode': 'namenode'
})
submission = Submission(None, properties=properties, oozie_id='test', fs=fs, jt=jt)
assert_equal(properties, submission.properties)
submission._update_properties('jtaddress', 'deployment-directory')
assert_equal(final_properties, submission.properties)
finally:
cluster.clear_caches()
for reset in finish:
reset()
def get_task_log(self, offset=0):
logs = []
attempt = self.task.job.job_attempts['jobAttempt'][-1]
log_link = attempt['logsLink']
# Get MR task logs
# Don't hack up the urls if they've been migrated to the job history server.
for cluster in YARN_CLUSTERS.get().itervalues():
if log_link.startswith(cluster.HISTORY_SERVER_API_URL.get()):
break
else:
if self.assignedContainerId:
log_link = log_link.replace(attempt['containerId'], self.assignedContainerId)
if hasattr(self, 'nodeHttpAddress'):
log_link = log_link.replace(attempt['nodeHttpAddress'].split(':')[0], self.nodeHttpAddress.split(':')[0])
for name in ('stdout', 'stderr', 'syslog'):
link = '/%s/' % name
params = {}
if int(offset) >= 0:
params['start'] = offset
try:
log_link = re.sub('job_[^/]+', self.id, log_link)
root = Resource(get_log_client(log_link), urlparse.urlsplit(log_link)[2], urlencode=False)
response = root.get(link, params=params)
log = html.fromstring(response, parser=html.HTMLParser()).xpath('/html/body/table/tbody/tr/td[2]')[0].text_content()
except Exception, e:
log = _('Failed to retrieve log: %s' % e)
try:
debug_info = '\nLog Link: %s' % log_link
debug_info += '\nHTML Response: %s' % response
LOGGER.error(debug_info)
except:
LOG.exception('failed to build debug info')
logs.append(log)
